If you’re a publisher, affiliate, ecommerce shop, or lead gen company and you send a newsletter, LLMs like ChatGPT, Perplexity, or Claude may be able to help you optimize them. This includes informative ones, email blasts for deals and sales, product announcements, and hybrid mixes. Just make sure your accounts are secure, the data will not get leaked, and always use the paid accounts with the LLMs.
Here are four ways LLMs can help you get better delivery or drive more conversions with your newsletters. These strategies and ideas can be taken multiple levels deeper, it all depends on how granular you want to go.
For each of these you’re going to need to download the relevant metrics including:
- Click through rates
- Open rates
- Bounce rates
- Unsubscribes
- Conversions ($ or # or both on separate lines)
- Subject lines
- Calls to action (style as in colors and buttons, wording, etc…)
Strategy 1: Sorting Your Database Based on Intent and Engagement
Instead of having to figure out which subscribers want more content vs. promotions, or sending and hoping for the best, let AI do the work for you. Start by downloading the last four or five results of email campaigns and use prompts to sort and group the lists. This can include topics the person is interested in, formats that create more clicks by email, etc…
If you have an informative newsletter with different sections like finance, marketing, etc… ask the LLM to sort the email addresses into groups based on click behavior from the last four or five emails. Here’s a potential sample prompt for ChatGPT.
“Please take the analysis of email addresses that clicked on these last X email campaigns and create separate lists including “clicks multiple types of content”, “clicks finance only”, “clicks marketing only”, “(insert others)” and only include the individual email address in one list.
Include the email addresses that click at least two different topics on multiple newsletters in the “multiple types of content” list and not in any of the others. The first section of the email which can be found in (insert how to identify it based on your template like the third H2 tag labeled “finance”) is finance, the second section is marketing identified by (insert the identifier), and the third is hiring trends. Now please label the lists to match the topic and place them in a spreadsheet with the first name and email address.”
Now you’ll have your newsletter subscribers broken out by what interests them. You can send the same newsletter but prioritize what goes on top of the format, or send niche specific ones so people only interested in marketing get marketing, and people that like all topics get a general one.
This works for conversion newsletters with deals, discounts, product categories, bonus points or miles, etc… Instead of using the topics, use the types of deals they want or types of products. It could be running shoes vs. designer stilettos, or gluten free recipes vs. decor items. For deal sites try by the type of store (cosmetics vs. sports), specific brands, and prompt to remove irrelevant brands based on click and conversion history to try and maximize earnings per email campaign.
Optimize the Bounce and Open Rate by ESP
The next way LLMs like ChatGPT, Perplexity, and Claude can be used are to pick the best subject line based on email service provider to help with delivery and open rates. This can be combined with the body style in the next section.
Start by uploading your email list and include the bounce rate, open rate, and spam box if provided. Please note that when spam box data is provided, it is normally estimates based on the tools own algorithms and does not mean it actually did go to spam. Look for open rate and click number declines as a way to determine you could have an inbox delivery issue.
“Please take the list I uploaded and sort the list by email service provider which is the word after the @ symbol for email service providers with more than 50 subscribers including gmail.com, yahoo.com, and hotmail.com. Now take the last five emails and evaluate which subject lines get the best delivery, open, and click through rates by email service provider. My new newsletter is about X and Y topic and will include A, B, and C. I’d like the theme to be 123. Please create a subject line for each email service provider that has the best chance of delivery, open rates, and click through rates, then separate the email addresses into individual lists. For the email service providers with less than 50 subscribers, place them in a combined list called combined newsletter subscribers and generate a subject line for them as well.”
Select Body Style Based on Users
Formatting the newsletter template using a grid, a large list, or sections with images and text with three across or a grid of four can be optimized with ChatGPT, Perplexity, and Claude using similar prompts. This is all based on the metrics you gather and enter including opens, click through rates, conversions (if you have the data), topics of interest, unsubscribes.
The ways to get the lists here can be based on layout (grids vs. lists), images vs. text only, combinations of the two, and most important are the metrics you’re looking for. The goal could be sending the same newsletter with multiple layouts based on how the subscribers engage, or creating a template for the complete list that has the best chances of opens and click throughs, and the least chance of getting unsubscribes or not deliverable issues. Then add the subject lines by having the LLM create them.
Coming Up With Subject Lines
LLMs can come up with subject lines for open rates, deliverability, and what to avoid by uploading the lists like above. Now enter what you’ll be including in the newsletter and what happens once the links are clicked. i.e. the person goes to the store and the page they’ll find, or the subscriber goes to your own website where they’ll learn about X, Y, and Z topics via articles while finding 123 which can be products or solutions for them.
“Please evaluate the last twenty newsletter subject lines and look for what they have in common in theme for the five that had the best open rates and least unsubscribes or complaints. I need a subject line for the entire list that has the least chance of getting unsubscribes and complaints. The theme of the newsletter is XYZ, I’ll be including 123 articles/brands, and once the user clicks through they’ll find D, E, and F on the landing pages. Please provide me with 5 subject line options under 55 characters and share why you think each one will be good.”
These same types of prompts can be done for figuring out which brands to feature in deal or sale email blasts as well as which topics are most interesting by month, quarter, or subscriber. As long as there is no legal issue with uploading the email data to the LLM system and asking for an output, you may be able to grow your engagement and revenue numbers by having AI help with your newsletter lists.
